Historical vegetation change in the Kananaskis Valley, Canadian Rockies
E. A. JOHNSONA N D G. I. FRYER
Division of Ecology, Department of Biology, and Kananaskis Centre for Environmental Research, University of Calgary,
Calgary, Alta, Canada T2N IN4
Received June 23, 1986
JOHNSON,
E. A., and FRYER,
G. I. 1987. Historical vegetation change in the Kananaskis Valley, Canadian Rockies. Can. J.
Bot. 65: 853-858.
This study compares the vegetation composition in the Kananaskis Valley from a forest survey in 1883 to another survey in
1972 and reconstructs the fire frequency for the period 1783- 1882 and the period 1883- 1972. A comparison of the 1883 to
1972 forest surveys using transition probabilities revealed that sites populated mostly by lodgepole pine (Pinus contorta) or
Englemann spruce (Picea Englemannii) tended to remain the same in both surveys. The fire reconstructions for the period
1730- 1972 showed no change in fire frequency after the beginning of European activity in 1883. Further, for the periods both
before and after 1883, the valley burned, on average, once every 150 years. The distribution of the fire sizes for the 100-year
period before 1883 was slightly larger than the period after 1883. It appears that in 1972 natural processes (site differences and
fire occurrence) still dominated the changes in the vegetation composition and age.
JOHNSON,
E. A., et FRYER,
G. I. 1987. Historical vegetation change in the Kananaskis Valley, Canadian Rockies. Can. J.
Bot. 65 : 853-858.
Dans la prCsente Ctude, la composition floristique de la VallCe de Kananaskis, d'aprks une Cvaluation forestikre faite en 1883
est comparCe k celle d'aprks une Cvaluation faite en 1972, et la frCquence des incendies pour la pCriode de 1783- 1882 et celle
de 1883- 1972 est Ctablie. Une comparison des Ctudes forestikres de 1883 et 1972, k l'aide de probabilitis de transition, a
dCmontrC que les sites peuplCs surtout de pin lodgepole (Pinus contorta) ou d'Cpinette bleue (Picea Engelmannii) avaient
tendance k demeurer inchangCs, et ce, dans les deux Ctudes. D'apres la reconstitution des incendies pour la pCriode
1730- 1972, il n'y a pas eu de changement dans la frCquence des feux aprks le dCbut de I'activitC europCenne en 1883. De
plus, pour les pCriodes tant avant qu'aprks 1883, la vallCe a CtC en feu, en moyenne, une fois tous les 150 ans. La distribution
de l'importance des feux pour la pCriode de cent ans avant 1883 Ctait 1Cgkrement plus grande que pour la pCriode aprks 1883. I1
semble qu'en 1972, des processus naturels (diffkrences des sites et la survenance des feux) dominaient encore les changements
dans la composition et l'ige de la vegetation.
[Traduit par la revue]
Introduction
We take it for granted that European impact has caused
serious changes in vegetation composition and age in North
America. How much European activities have affected vegetation relative to natural processes (including Indians' utilization) is, however, the question which must be determined
(Kline and Cottam 1979). For example, Byrne (1964) and
Nelson and Byrne (1966) examined the changes in vegetation
in Banff National Park brought about by the railroad, lumbering, mining, and tourism after 1886. Using old photographs
and contemporary accounts, they present seemingly convincing evidence that Europeans had significantly increased the
occurrence of fires from 1886 to 1910 and probably changed
the forest. The occurrence of the fires is a fact that can be
documented (White 1985). What is difficult, as Nelson and
Byrne (1966) point out, is to evaluate if European activities
changed the composition and age distribution of the forests,
and whether or not the frequency of fires increased during this
period.
The objective of this study is to compare the vegetation
composition and age structure in the Kananaskis Valley in
1883, before European utilization began, and in 1972, after
logging, mining, and hydroelectric development had occurred.
We use techniques which reconstruct the regional forest
composition (timber surveys) and forest age (fire history) in
the valley. In contrast to old photographs, diaries, and remembrances (Forman and Russell 1983), timber surveys and fire
histories are systematic and allow a greater part of the landscape to be reconstructed for specific traits. These results
should reflect the relative importance of European activity
versus natural processes in determining the composition and
age of the vegetation in 1972.
Printed in Canada / lrnprirnt au Canada
The Kananaskis Valley is approximately 100 krn west of
Calgary, Alberta, and directly southeast of Banff National
Park. It is the first major intermountain valley in the Front
Range of the Canadian Rockies and the major tributary of the
Bow River. The valley is approximately 55 km long, running
NNW-SSE in weak Mesozoic clastic rocks. The ridges,
which dip to WSW, are subparallel thrust sheets and fault
blocks composed of resistant Paleozoic carbonate rock. The
valley was glaciated during the Pleistocene. Precipitation
varies with elevation, increasing about 20 mrn for every
100 rn of elevation gain; average annual precipitation at the
Kananaskis Lakes (elevation, 1670 rn) is 610 rnm (Eastern
Rockies Forest Conservation Board 1968). Forest fires can
occur from April to October. The vegetation consists of a
lower zone (1000- 1400 m) of lodgepole pine (Pinus contorta
Dougl. var. latifolia Engelm.), Engelmann spruce (Picea
Engelrnannii Parry), and some subalpine fir (Abies lasiocalpa
(Hook.) Nutt.) and an upper (1400-2100 m) zone of Engelmann spruce and subalpine fir.
Methods
Nature of European impact from 1883
Unpublished historical records of logging in the Kananaskis Valley
came from the Eau Claire & Bow River Lumber Co. papers and the
James Walker papers in the Glenbow-Alberta Institute Archives,
Calgary. Unpublished information on coal mining and mineral claims
from the Kananaskis Exploration and Development Co., Calgary, and
the George W. Pocaterra papers in the Glenbow-Alberta Institute
Archives, Calgary. Hydroelectric development information came
from Calgary Power Ltd., Calgary.
Comparison of 1883 to 1972 timber surveys
Louis Stewart surveyed the timber resources in the Kananaskis,
CAN. J.
854
BOT. VOL. 65,
1987
TABLE1. The transition probability matrix between the composition-size categories in 1883 and the composition-size categories
in 1972
1883
timber
survey
1972 Timber Survey
Pine,
Pine,
Pine,
Pinespruce,
Pinespruce,
Pinespruce,
Spruce,
Spruce,,
Spruce,
Spruce aspen
Burned
Pine,
Pine,
Pine,
Pine-spruce,
Pine-spruce,
Pine -spruce,
Spruce,
Spruce,
Spruce,
Spruce -aspen
Burned
NOTE:S , small; m, medium; I, large.
Spray, and Bow Valleys for the federal government in the summer of
1883, subdividing the area into timber berths and evaluating the
timber size and composition (Eau Claire & Bow Lumber Co. papers,
Glenbow-Alberta Archives). Stewart recorded the tree size, species,
bums, and blow-downs, etc., at 600-m intervals, and for approximately 1300 m on both sides of the Kananaskis River to the
Kananaskis Lakes. The Alberta Forest Service (1975) produced a
forest cover survey based on 1972 aerial photographs of the same area
of the Kananaskis as Stewart's 1883 survey. The information on
forest composition and size in both the 1883 and 1972 surveys was
divided into eleven composition-size categories based on Stewart's
categorization of the vegetation (see Table 1 for categories).
To analyze the compositional and size changes between surveys,
transition probabilities were calculated in the following manner
(cf. Kemeny and Snell 1960). First, the number of transitions from
one composition-size category to any other were counted. Next, the
transition probabilities were calculated by pij = nijlFnij,where nij was
the number of transitions from composition-sizk category i to j
(i = 1883 survey and j = 1972 survey). Note that Cpij equals 1,
i.e., a composition-size category i moves with certainty to another
composition-size category, including itself. To remove the transitions expected by chance from pij, we used rij = njlN, where nj was
the number of occurrences of composition-size category j, and N is
the total number of occurrences of all composition-size categories.
By subtracting pij from rij, the transitions not due to chance are given
on a scale from -1 to + l .
Fire history
Our objective was to reconstruct (i) the size and distribution of fires
occumng in the periods 1783 to 1882 and 1883 to 1972 and (ii) determine the time-since-fire distribution (Johnson and Van Wagner 1985)
as of 1972 for the Valley. The method for collecting the data for these
reconstructions was as follows. The boundaries of vegetation which
appeared to be of different fire ages were delineated on aerial photographs taken in 1972. Using these boundaries as first approximations
of a stand origin map (cf. Heinselman 1973), 900 point samples of
5 ha (cf. Johnson and Van Wagner 1985) were placed so that one or
more samples were present in each vegetational area of different fire
age. In each of these samples, the fire history was reconstructed using
fire scars and tree ages. More samples were added when vegetation of
different fire ages were recognized in the field. From the fire chronologies (cf. McBride 1983) of the samples, the extent of past fires were
reconstructed and the final stand origin map for 1972 drawn. Areas of
both past fires and stand origin years were measured on a 1 : 50 000
scale map using a Ziess MOP digitizer. The time-since-fire distribution was plotted on Weibull probability paper (Johnson and Van
Wagner 1985) and the average fire interval was determined graphically (King 197 1).
Results and discussion
The nature of European impact from 1883
The first Europeans known to have traversed the Kananaskis
Valley were a group of settlers led by James Sinclair in August
1854. Sinclair found the North Kananaskis Pass route difficult
for wagons and settlers because of downed fire-killed timber
(Oltmann 1976). The Palliser Expedition led by Captain John
Palliser explored the valley in 1858 as a route for the railroad through the mountains. The Kananaskis passes were subsequently rejected as a usable pass for the railroad (Spry 1968).
After 1887, all commercial travel was directed through the
Banff Valley and Kicking Horse Pass.
Significant European utilization began with a tour of the
Kananaskis, Spray and Bow valleys in 1883 by I. K. Kerr of
Calgary, K. MacFee of Ottawa, and two business associates
from Eau Claire, Wisconsin. Their objective was to find
marketable timber for a sawmill to be located in the newly
founded town of Calgary. In the Kananaskis Valley, they bid
on the leases for four timber berths (Fig. 1). The berths were
leased under the Dominion Lands Act of 1879 which precluded
forested lands from sale or homestead entry (Martin 1938).
In 1886, the Eau Claire & Bow Lumber Co. began the first
organized logging operation in the Kananaskis Valley. Engelmann spruce, lodgepole pine, and some subalpine fir were the
trees sought. Spruce was preferred for saw timber because of
its larger size, but pine was also harvested as it was straight
and excellent pole lumber (Eau Claire & Bow Lumber Co.
papers, unpublished Glenbow-Alberta Institute Archives). In
the years before 1920, the diameter limits were 8 to 16 in.
(1 in. = 25.4 mm) and lengths 12, 14, 16 ft. (1 ft =
0.3048 m). Other small logging operations in 1932- 1933 and
1942- 1947 salvaged fire-killed timber (Oltmann 1976).
Until the 1940's, when chain saws were introduced, loggers
felled trees using axes and crosscut saws. After sawyers
trimmed limbs and sawed the boles into proper lengths,
skidders, using horses, dragged logs directly to rollways on the
river or to skidways where they would be loaded on sleds for
transport to the river. Roads, skidways, and sleds were only
used in more level parts of the main valley or in large side
valleys such as Ribbon and Evan-Thomas creeks (berth H);
steep slopes were not usually logged because of the difficulty
of removing logs by horse. Trucks were used for the first time
when salvaging timber after the 1936 fire in berth H.
JOHNSON AND FRYER
FIG. 1. The original map of L. B. Stewart, showing the location of the timber berths surveyed by him in 1883. Berths G, H, I, and J are the
berths in the Kananaskis Valley leased by the Eau Claire & Bow Lumber Company, and used in this study. Source: Glenbow-Alberta Institute,
Calgary, Alberta. (Printed with permission.)
FIG.3. The total area burnt in the Kananaskis Valley for the period
1783- 1882 and the period 1883- 1972.
FIG. 2. Graphic representation of all transitions greater than
expected by chance. The circles are the composition-size categories,
in which the abbreviations are the same as in Table 1. The arrows
represent the transitions between the composition-size categories.
The probabilities of the transitions are given next to each arrow.
Logs were removed to the Calgary mill by water, the log
drives taking about 2 months. Actual logging took place
mostly in the winter. The first spring log drive was in 1887 and
the last drive in 1944, 1 year before the Eau Claire & Bow
Lumber Co. went out of business. Although detailed records of
where and when cutting occurred are lacking, we may surmise
that the first cutting was largely done in berth I, since the first
winter camp was located there in 1886. After the turn of the
century, a second winter camp was constructed in berth H, and
presumably, this is where the next cutting was done. In 1920,
parts of berths H, I, and J burned. The depression of the 1930's
reduced logging activities and a large fire in 1936 killed most
of the trees in berths G and H. Most of the logging from 1936
to 1944 was the salvage of fire-killed timber in berth H. Berth J
was never logged because a large log jam prevented the river
from being used.
Coal was discovered in the Kananaskis Valley at the turn of
the century, and by 1907, George Pocatem had filed claims on
856
CAN. J . BOT. VOL. 65, 1987
FIG.4. The reconstruction of individual fires in the Kananaskis Valley from 1783 to 1972.
most of the promising areas (G. Pocaterra papers, unpublished
Glenbow-Alberta Institute Archives). However, with the
exception of the Mount Allan (Ribbon Creek) coal claim, all
were unexploited. The Mount Allan mine operated from 1947
(strip mine, 1948 underground) to 1952. It was always a small
operation and used timber primarily from around the mine and
Ribbon Creek. Most of the area near the mine had been burned
by the 1936 fire. Other mineral deposits (e.g., gypsum) were
explored, but were never economically viable (Oltmann 1976).
Probably the most lasting European impact on the Kananaskis Valley has been the construction of Bamer Lake and the
expansion of the upper lakes for hydroelectric development.
The first hydroelectric dam on the upper Kananaskis Lake was
a wooden spillway in 1932. Bamer Lake dam was constructed
in 1945. The timber from the flooded area was cut and burned,
not being big enough to sell. The lower Kananaskis Lake dam
was built in 1954. Except for the dam and related control structures, however, there has been no other impact on the vegetation from hydroelectric development. In total, this development flooded less than 1% of the Valley.
Farming and ranching never became established in the
Valley owing not only to the Dominion Lands Act of 1879, but
also to the longer winters, deep snow, and lack of forage,
when compared with the adjacent foothills.
Comparison of the 1883 to 1972 timber surveys
The transition probability matrix in Table 1 gives an estimate
of the probability of a single site starting at some initial composition-size category in 1883 and amving at some final
composition-size category in 1972. Table 1 gives all the transitions. By removing the transitions greater than expected by
chance ( > 0) from Table 1, a pattern becomes clearer (Fig. 2).
Two clusters of transitions are evident in Fig. 2., one consisting of pine categories and the other of spruce categories.
The pine and spruce clusters indicate that sites populated by
pine did not shift to spruce, and vice versa. We believe that
these clusters in the forest surveys reflect differences between
sites that favor the establishment and growth of lodgepole pine
or Engelmann spruce. The transitions within the pine and
spruce categories reflect the specific disturbance history of the
sites. For example, a site populated by medium spruce in 1883
could change to either small or large spruce, depending upon
the time after 1883 in which disturbance occurred. If the area
was burned in 1891 and not disturbed again before the 1972
survey, it could have changed into a large spruce category. As
this example indicates, it is important not to interpret the
changes within clusters as simply succession, since most areas,
as we shall see, would have burned between the surveys.
There are three transitions between the two clusters that have
probabilities larger than 0.10, which require an explanation
because of the change in species composition. The first transition from the small pine -spruce to small pine between surveys
can be explained by the difference in minimum seed-bearing
age of the two species (and no nearby surviving seed-bearing
spruce trees). Pine can produce cones between 10 and 15 years
of age yet generally do not have serotinous cones until
20 years of age (personal observation). Spruce, on the other
hand, do not produce cones until 16 to 25 years of age. Therefore, fire or other disturbance(s) between the two surveys
would probably find the pine able to contribute to the next
generation, while spruce at the same age could not. The second
transition, from the small pine-spruce to the medium pine
categories, could be due to similar reasons. Finally, the transition from the small spruce to the spruce - aspen category could
be due to more than one disturbance, providing advantage to
the vegetatively reproducing aspen over sexually reproducing
spruce.
Fire history
Forest fires have burned most of the valley bottom at least
once since 1883 (Fig. 3), with four fires (1891, 1904, 1920,
and 1936) accounting for most of the area burned (Fig. 4). The
period 1783 to 1882 shows a similar fire pattern (Fig. 3);
again, most of the valley bottom burned, and six large fires
(1803, 1853, 1858, 1865, 1870, and 1881) accounted for most
of the area burned (Fig. 4). Comparison of the fire-size distributions (Fig. 5) between the two periods indicates that the
period 1783- 1882 had slightly larger fires than 1883- 1972.
It is impossible with the present state of our knowledge to
determine if this is the result of the fire protection required by
the Dominion Lands Act of 1879. Fire ~rotectionwould have
been more effective against fires with low average spread rates
and thus, smaller eventual size. Further, the difference
between the two periods is not great, which suggests that fire
protection did not seriously distort the fire-size distribution.
The time-since-fire distribution (Fig. 6) gives the probability
of surviving without a fire in the Kananaskis Valley. The break
in the line at 1730 may indicate a change in fire frequency.
-
JOHNSON AND FRYER
Rank f r o m largest to smallest
FIG. 5. Distribution of relative fire sizes for the periods 1783 to
1882 ( 0 )and 1883 to 1972 (0).
For our concern here, the period 1972 to 1730 has a similar
slope which indicates an average fire interval of 150 years.
Therefore, the European utilization of the Kananaskis Valley
after 1883 does not seem to have changed the fire frequency
from the previous 153 years.
Conclusion
Our results indicate that although Europeans have utilized
the Kananaskis Valley's timber, mineral, and water resources
since 1883, the forest composition and age structure have not
been significantly altered. This appears to be due to the type of
European land use which allowed the topography and fire
regimes to continue to determine the vegetation pattern.
Land use in the Kananaskis Valley was determined by the
Dominion Lands Act of 1879 which barred the sale of forested
lands for homesteading. The snowy winters and short growing
season make the valley unsuitable for ranching and farming,
but the Dominion Lands Act of 1879 prevented any attempts at
ranching or farming. If these attempts had been made, forests
could have been converted into grazing land and open fields,
which would have had a more lasting impact on the vegetation.
Logging was the most significant activity in the Kananskis
Valley. Although the Dominion Lands Act provided for fire
protection, the fire history would seem to indicate no change in
fire frequency as a result of the increasingly effective fire
protection in recent years. The majority of the areas logged
were either subsequently burned or logging was of already
burned timber. This greatly reduced the impact of logging in
the area.
Logging was done only in the winter, and the sleds utilized
for carrying timber did not require permanent roads. Further,
both logging and mining did not encourage permanent settlements. The railroad was built through the Banff Valley instead
of the Kananaskis Valley, so that there was never a persistent
flow of traffic through the valley. Therefore, for all of these
reasons, European impact was greatly reduced from what it
potentially could have been.
There are limitations to the results we have obtained. The
study is limited to the forest composition and size changes in
the time period and the area surveyed. Although there may be
other characteristics which changed with European activity,
we have not investigated these. We have chosen the vegetation
in 1783 to 1882 as the standard against which we measure
STAND AGE (years)
FIG. 6. The time since fire distribution for the periods 1972 to 1589.
The lines of best fit for these points were drawn using Fenill's median
regression method (King 197 1).
European impact. This period still had some Indian and
European activity in the valley, although limited, and by no
means can the vegetation be considered to be influenced only
by site differences and lightning fire occurrences.
Acknowledgements
We thank M. C. Kellman, R. C. Oltmann, W. F. J. Parsons,
and J. W. Sheard for thoughtful comments on the manuscript.
T. Leonard helped analyze the timber surveys, and C. Larsen
and H. Friebe collected the fire history data. The GlenbowAlberta Institute helped locate historical records and photographs. The study was supported by a Natural Sciences and
Engineering Research Council of Canada operating grant and
Alberta Environment Trust grant to E. A. J.
ALBERTAFORESTSERVICE.1975. Forest Management Unit map.
Department of Lands and Forest, Government of Alberta.
BYRNE,A. R. 1964. Man and landscape change in the Banff National
Park area before 1911. M.A. thesis, University of Calgary,
Calgary, Alta.
EASTERNROCKIESFOREST CONSERVATION
BOARD.1968. Some
climate data for the east slopes of the Rockies in Alberta. Eastern
Rockies Forest Conservation Board Management Rep. no. 2.
Eastern Rockies Forest Conservation Board, Calgary, Alta.
FORMAN,R. T. T., and RUSSELL,E. W. B. 1983. Evaluation of
historical data in ecology. Bull. Ecol. Soc. Am. 64: 5-7.
HEINSELMANN,
M. L. 1973. Fire in the virgin forests of the Boundary
Waters Canoe Area, Minnesota. Quat. Res. (N.Y.), 3: 329 -382.
JOHNSON,
E. A., and VANWAGNER,
C. E. 1985. The theory and use
of two fire history models. Can. J. For. Res. 15: 214-220.
KEMENY,J. G., and SNELL,J. L. 1960. Finite Markov Chains.
D. van Nostrand Co. Inc., Princeton, NJ.
KING,J. R. 1971. Probability charts for decision making. Industrial
Press, New York.
KLINE,V. M., and COTTAM,G. 1979. Vegetation responses to
climate and fire in the driftless area of Wisconsin. Ecology, 60:
861 -868.
MARTIN,C. 1938. "Dominion Lands" policy. In Canadian frontiers
of settlement. Vol. 2. Edited by W. A. Mackintosh and W. L. G.
Joerg. Macmillan Co. of Canada Ltd., Toronto.
MCBRIDE,J. R. 1983. Analysis of tree rings and fire scars to establish
fire history. Tree-Ring Bull. 43: 51 -67.
NELSON,J. G., and BYRNE,A. R. 1966. Fires, floods and national
858
CAN. J. BOT.
parks in the Bow Valley, Alberta. Geogr. Rev. 56: 226-238.
OLTMANN,
R. C. 1976. The valley of rumours . . . the Kananaskis.
Ribbon Creek Publishing Company, Seebe, Alta.
SPRY,I. M. (Editor). 1968. The papers of the Palliser Expedition
1857- 1860. The Champlain Society, Toronto.
VOL. 6 5 ,
1987
WHITE, C. A. 1985. Wildland fires in Banff National Park
1880-1980. Occasional Paper no. 3. National Parks Branch,
Parks Canada, Environment Canada.